WO1998039943A1

WO1998039943A1 - Method of measuring traffic densities

Info

Publication number: WO1998039943A1
Application number: PCT/FI1998/000150
Authority: WO
Inventors: Petri Jolma
Original assignee: Nokia Telecommunications Oy
Priority date: 1997-02-24
Filing date: 1998-02-19
Publication date: 1998-09-11
Also published as: JP2000509947A; NO984941D0; CN1217859A; FI970760A0; FI970760L; AU6216898A; NO984941L; AU732025B2; EP0935897A1; FI107500B

Abstract

When increasing the capacity of a cellular radio network, a second base transceiver station of less power is often installed within the original cell to serve traffic hot spots in the original cell. The final outcome is strongly dependent on how well the location of the traffic hot spot and the number of mobile stations can be defined. In the invention, the location of traffic hot spots and the number of mobile stations are measured by locating such a test transmitter in the cell which will cause interference to the signal received by mobile stations. Any changes caused by the interference can be detected in the quality reporting of mobile stations.

Description

Method of measuring traffic densities

Field of the invention

This invention concerns a method of measuring traffic densities in a cellular radio system.

Background of the invention

Figure 1 shows the structure of a cellular radio system. The system comprises an MSC (Mobile Services Switching Centre), a BSC (Base Station Controller), a BTS (Base Transceiver Station), an MS (Mobile Station) and an NMS (Network Management System). The network typically comprises several interconnected mobile services switching centres MSC, only one of which is shown in the figure for the sake of clarity. The mobile station system is connected to a fixed telephone network, e.g. a PSTN (Public Switched Telephone Network) or an ISDN (Integrated Services Digital Network), by way of the mobile services switching centre MSC. Typically, several base station controllers BSC subordinated hierarchically to the MSC are connected to each mobile services switching centre MSC through a so-called A interface. Also typically, several base transceiver stations BTS hierarchically under the BSC are connected to each base station controller BSC through a so-called A-bis interface. Through channels of a radio path passing over the air interface the base transceiver stations may form connections with mobile stations MS. To keep it simple, the figure shows just one base station controller BSC, one base transceiver station BTS and one mobile station MS. The network management system NMS is used to collect information on the network state and to manage the operation of network elements, e.g. by changing the network configuration. A system of this type is e.g. the GSM system which is described more extensively in a book by Michel Mouly & Marie-Bernadette Pautet titled "The GSM System for Mobile Communications". When designing a cellular radio network the area where radio coverage is desired is divided into cells, each one of which is served by a base transceiver station. The cell size is determined as a compromise between the desired network capacity and the price of the investment. Due to the high cost of base transceiver station investment, designers aim at keeping the number of base transceiver stations as low as possible. The number of channels used in individual cells and this way the capacity of the entire network are limited by interference between connections which are caused to each other by connections working in adjacent cells on the same channel or on an adjacent channel. Of course, mutual interference is higher the more closely to one another the connections are working. Due to interference, a channel can not be used simultaneously in cells which are closer to each other than the so- called reuse distance. In present day practice, channels are typically reused either in every ninth cell or in every 12^th cell.

When using low-power and thus short-range base transceiver stations, the same channels may be used geographically closer to each other than when using base transceiver stations of higher power, whereby a higher total capacity is achieved for the system. It is thus possible to increase network capacity by using base transceiver stations with as short a range as possible and forming small cells. On the other hand, due to the high costs of base transceiver station investments, the number of base transceiver stations must be minimised.

The number of channels required in a cell is estimated by taking into account the estimated number of users within the service area of the cell, their expected call activity and the allowed barring. The required number of channels is allocated so that channels allocated from different base transceiver stations will not cause too much interference to one another. The present typical allocation of channels for use by base transceiver stations is by utilising FCA (Fixed Channel Allocation) demanding fixed network design. Other possible allocation methods are e.g. DCA (Dynamic Channel Allocation) and their combination HCA (Hybrid Channel Allocation). As the use of mobile stations is increasing along with applications demanding wide band widths, such as e.g. multimedia applications, the estimates made of the number of channels required in cells at the network construction stage are often too small. Hereby mobile stations in the cells begin experiencing such barring of calls, especially in the rush hour, which exceeds the dimensioned barring. This can be established e.g. by taking normal network management steps to find that the cell filling degree is very high during rush hours, e.g. an average of 90 per cent of all channels in use. It is also found that all channels of the cell are busy at the same time during the rush hour for a considerable part of the time. In addition or alternatively it may be found that due to excessive loading of the cell a high number of channel requests from mobile stations must be dismissed. The conclusion can be drawn from these findings that the cell capacity must be increased to make the barring noticed by users acceptable.

The easiest way to increase cell capacity would be to increase channels available to the base transceiver station by using new frequencies. However, because the introduction of new frequencies in a cell would typically raise the interference level of the network too much, this can not be done in several cases. If there are only a few such cells in the network , where the traffic density is too high in relation to the capacity, then dynamic channel allocation DCA may also be used or borrowing of channels from adjacent cells (see e.g. T -S.P. Yum and W.-S. Wong: Hot-Spot Traffic Relief in Cellular Systems, IEEE Journal on Selected Areas in Communications, vol. 11 , no. 6, August 1993). However, these methods limit the use of channels elsewhere in the network and make the network more complex. The network load may even become so high that the channels which can be used with the original cell structure are not enough even when optimally used. The only remaining alternative is then to change the cellular structure of the network.

A generally used method of increasing capacity is to install a new base transceiver station with lower transmission capacity in the network. Hereby on the one hand, its range or cell size is smaller, whereby less users remain to be served by the individual cell, and on the other hand it causes less interference to nearby base transceiver stations, whereby the frequencies used by the base transceiver station may be reused closer to one another. Since the total capacity of the cellular radio network can be increased in this way, the present trend is to go in for ever smaller so-called microcells and picocells. On the other hand, one must know how to locate these smaller cells to optimise efficiency so that they will relieve the load of the original base transceiver station as effectively as possible.

Traffic density is never evenly distributed in the cell, but there are such traffic hot spots where the call traffic is denser than the average. Figure 2 shows a base transceiver station and its service area, where a traffic hot spot occurs. In these traffic hot spots a great part of the total traffic in the cell is in a small part of the cell. Such areas may be e.g. railway stations, street crossings and sport stadiums. When looking for a site for a new base transceiver station with lower power, the most sensible thing to do is to look for the cell's traffic hot spot or spots and to locate the new base transceiver station to serve such traffic hot spot. Information must be obtained on the location of traffic hot spots and on the number of mobile stations in them so that the new base transceiver station can be located in an optimum place. However, the problem is that mobile stations in the network can not be localised at present with greater accuracy than one cell. When trying to find the optimum location for a new base transceiver station, no great advantage is achieved by determining the location of mobile stations with one-cell precision. Due to this state-of-the-art problem, expensive base transceiver station investments must often be made only based on the assumed geographical distribution of traffic. The present invention aims at eliminating or at least at reducing this state-of-the-art defect. This objective is achieved with the method defined in the appended independent claims.

Brief description of the invention In the method a test transmitter is located in the network which transmits a signal causing a change of a known type in the signal received by mobile stations. The mobile stations perform measurements of connection quality in their normal manner and send the results of their measurements to the network in their measurement report. By comparing the measurement reports sent by mobile stations to the network with the known test signal it is possible to deduce which mobile stations are within the service area of the test transmitter.

It is advantageous for the practical embodiment to use a periodically working pulsed carrier wave as test signal, whereby when the pulse is on, a brief but surely noticeable change is obtained in the signal received by the mobile station. With such a signal, interference can be caused for mobile stations located within the service range of the test transmitter, e.g. in every hundredth bit, which will not noticeably impair the call quality noticed by the user, but which will suffice to calculate a quality value determined for the signal received by the mobile station. Hereby it is noticed in the measurement reports of all mobile stations in the test transmitter's service area that quality values follow the periodicity of the test signal.

Brief description of the drawings The invention will be described more closely in the following referring to the appended drawings, wherein Figure 1 shows the known hierarchical structure of a cellular radio network;

Figure 2 shows a base transceiver station and its service area where a traffic hot spot occurs;

Figure 3 shows a base transceiver station and its service area where a traffic hot spot occurs and where a test transmitter according to the invention is located;

Figure 4a shows an example of a test pulse sent by the test transmitter as a function of time;

Figure 4b shows a quality measurement report sent by a mobile station in the test transmitter's service area as a function of time; and

Figure 4c shows a quality measurement report sent by a mobile station outside the test transmitter's service area as a function of time.

Detailed description of the invention The operation of the method according to the invention is studied in the following using a GSM system as an example. Figure 3 shows a situation where the traffic hot spot of the base transceiver station in Figure 2 is searched with the method according to the invention using a test transmitter. The test transmitter is placed in the assumed traffic hot spot in Figure 3. The test transmitter is intended to send an interference signal the effect of which on the signal received by the mobile stations can be noticed. Based on these observations it is possible to determine whether the mobile station is within the service area of the test transmitter. Measurements are preferably made during the rush hour. The test transmitter is cheap and easy to manufacture. The transmitter need only be able to generate an interference signal of the desired shape and modulated for downlink frequencies, that is, for use from the base transceiver station of the cellular radio network towards the mobile station, such as e.g. the periodic pulse signal according to Figure 4. The transmitter may be entirely independent of the cellular radio network and it may e.g. operate with an in-built battery.

The test transmitter is located in the assumed place of a microcell and it is used for sending e.g. a signal according to Figure 4a having 10 microsecond interference pulses with intervals of 3 milliseconds periodically so that the signal is always on for 2 seconds and off for 2 seconds. The pulses interfere with the signal of the base transceiver station and cause an increase in the signal's bit error ratio. Hereby the quality of the signal received by telephones in the transmitter's coverage area is a little poorer when the signal is on than when the signal is off. However, since the pulse ratio is quite low, the interference will not cause any excessive attenuation in the signal received by the mobile stations.

Due to the channel coding used on the radio channel, the mobile station user will probably not even notice the impairment in quality caused by the test signal, although the impairment from the quality measured by the mobile station is noticeable. Channel coding will be described briefly in the following.

Due to interference, noise and drop-outs occurring on the radio path, data to be transferred is not transferred as such between the base transceiver station and the mobile station, but channel coding is first performed on the data. The purpose of channel coding is on the one hand to make information transfer better tolerate transfer disturbances and on the other hand to detect transfer errors. At the receiving end decoding may be performed e.g. using the known Viterbi decoder.

E.g. in a GSM system, the information to be transferred may be divided roughly into speech and data traffic, each of which is transferred in frame shape. Due to the different nature of speech and data traffic, different coding methods are used with them to achieve the best result possible.

In data traffic, block coding is first done on the data frame to be transferred, where check bits are added to the frame for detecting errors. Convolutional coding is done next, where n bit is always described as 2n bit, whereby an error correction possibility is achieved. Finally, the frame is also interleaved, which means changing the mutual order of bits. Since errors are typically short bursts of a few bits, interleaving achieves that bit errors occurring in the transfer will not occur in successive bits. On the other hand, the error correction made possible by convolutional coding works best with individual bit errors, so the error correction algorithm is made to work better with the aid of interleaving.

In speech traffic, the bits of the speech signal coded e.g. in accordance with specification GSM 06.10 are divided according to their importance into three classes, 1a, 1b and 2 respectively. Block coding is first performed on the bits of class 1a which are of major importance to the quality of the transferred speech. Convolutional coding is done on block-coded 1a bits and on non-coded 1b bits and the class 2 bits not coded before interleaving are added to the result of the convolutional coding.

A considerable bit error tolerance is achieved for the channel through channel coding. According to GSM specification GSM 05.07, a bit error ratio of approximately 3 % in channel coded data transferred on the channel will not yet cause any errors in the user's data to be transferred.

According to the above, it is possible to bring about changes in the bit error ratio noticed by the mobile station so that no errors at all will be caused in the data proper to be transferred. Mobile stations in the active state perform their measurements monitoring the signal quality quite normally without any changes required by the method according to the invention. In accordance with the state of the art they also define the bit error ratio of the signal they receive. The bit error ratio can be defined in many different ways. One useable method is to re-code the received and decoded data and to compare the result thus obtained with the original received signal.

The mobile station reports the results of its quality measurements to the base transceiver station on an SACCH (Slow Associated Control Channel) for the individual connection. Due to the limitation of the spectrum in use, the bit error ratio is not sent as such over the radio path, but the quality values are divided into eight rougher quality classes in accordance with the appended table.

By comparing the value of the 3 % error ratio shown in GSM 05.07 with the values of the above table, one finds that the data signal even in quality class 4 is free of errors in practice owing to channel coding. At the present time, in well designed GSM networks a majority of the quality received by the telephone obtains the value RXQUAL_0 indicating the best quality.

Through their measurements of channel quality, mobile stations located within the service area of the test transmitter detect the decline in bit error ratio of the signal which is caused by the test signal. The magnitude of decline in the bit error ratio can be adjusted e.g. by pulse ratio selection. E.g. by choosing a pulse ratio of 100, interference will be caused only to every hundredth bit and the increase in bit error ratio caused by the test signal will be approximately 0.5 %. It can be seen from the table above that this corresponds with quality class RXQUAL_3. However, even then the decline is so small that the quality of the call will not decline noticeably. Thus, the mobile station user need not necessarily notice the existence of the test signal. However, the quality value stated by the telephone in its measurement report will drop from the value RXQUALJ) to the value RXQUAL_3. Analogously, by using the pulse ratio 300 of our example, that is, e.g. a pulse of 10 microseconds at intervals of 3 milliseconds, the increase in bit error ratio will be about 0.2 % and the quality class will be RXQUAL_1.

In the case of our example, the test signal is periodic in accordance with Figure 4a, that is, the pulse signal is on for 2 seconds and off for 2 seconds. Of course, through their measurements mobile stations within the service area of the test transmitter will notice the decline in quality only when the pulse signal is on. For this reason, in their measurement reports the mobile stations within the service area of the test transmitter always report the value RXQUAL_0 according to Figure 4b for a time of two seconds and the value RXQUAL_1 for the following two seconds. Correspondingly, those mobile stations which are not within the service area of the test transmitter will not receive any periodic interference from the test transmitter's signal, but they will report uniform quality values in accordance with Figure 4c.

The mobile stations send their measurement results to the network normally in accordance with specification GSM 04.08. The network uses the received measurement reports for controlling the transmission power and for making any decisions to change channel, which functions are typically located in base station controller BSC. In the method according to the invention these entirely normal quality reports are also examined for such dependencies on the test signal, on the basis of which it is established whether the mobile station is within the service area of the test transmitter. When studying quality measurement results obtained from measurement reports sent by mobile stations on the SACCH channel it is found that the quality report of some mobile stations are always periodically at value 0 for 2 seconds and at value 1 for 2 seconds in accordance with Figure 4b. The conclusion can be drawn from this that the said mobile stations are located within the coverage area of the test transmitter and, which is more essential from the viewpoint of traffic hot spot measuring, it can thus be concluded how many mobile stations are located within the coverage area of the test transmitter. If it is found by analysing measurement results that a sufficient number of the mobile stations located in the cell is not within the coverage area of the test transmitter, no traffic hot spot has yet been found. The test transmitter is then moved to the next place. The measurement is repeated until such a place is found where the measurements indicate that it is most sensible to locate a new base transceiver station. It should be noted that the method does not give any direct information on the geographical area of a traffic concentration, but such information on the number of mobile stations located within the transmitter's service area which takes drop-out and other characteristics of the radio channel into account. Under these circumstances, measurement information is obtained on the number of mobile stations directly serviceable by the base transceiver station, if the aerial of the base transceiver station is located in the test transmitter's place.

An examination of dependencies of measurement results on the test signal can be done e.g. through software in that network element which processes the measurement report even otherwise. E.g. in a GSM system this element is the base station controller BSC. Alternatively, the function can be performed e.g. with a physical analyser and a connected control computer.

If processing is done in a network element where the measurement reports of mobile stations are processed even otherwise, such as e.g. in the base station controller of a GSM system, no physical equipment will be required for processing of results, but this can be done through software. Measurement reports are processed for doing power control and channel change decisions for the connection between base transceiver station and mobile station over the radio interface, whereby the units in charge of these functions must in any case study the time series formed by the successive quality reports of the individual mobile station. Under these circumstances, such a function may be added directly to the said units which is used for determining the dependence of quality reports on the test signal. However, since traffic hot spots are not always measured in practice, but only when considering the location of a new base transceiver station, it is preferable at least in some cases to implement the function in a separate plug-in unit which can be connected temporarily to the base station controller. It must be noted, however, that the manner in which analysis of quality measurement results is implemented is not essential to the main idea of the invention.

The physical analyser could be e.g. a device monitoring the so- called A-bis interface between base transceiver station and base station controller (see Figure 1) and storing quality reports of different mobile stations which it picks up from the interface. Correspondingly, the device could be located directly to monitor traffic at the air interface and to pick up quality reports of individual mobile stations directly from this interface. When determining the transmission power of the test transmitter the difference between service area and dominance area must be taken into account. The service area of a transmitter or base transceiver station is the area wherein the mobile station can detect the transmitter signal. This signal causes interference to another signal transferred at the same frequency, although it is not necessarily more powerful than the said other signal. The service areas of adjacent base transceiver stations are always slightly overlapping. The base transceiver station's dominance area again is the area wherein the base transceiver station's signal is most powerful. Mobile stations always seek connection with that base transceiver station in the dominance area of which they are located. Unlike the service areas, the dominance areas of adjacent base transceiver stations are not overlapping.

The test signal can be chosen to have a band corresponding to a traffic frequency used in operation. If frequency jumping is used in the network, it is preferable to use a test signal having so wide a band that it will cover the whole frequency area used for frequency jumping in the cell in question.

The test signal may cause interference also to other such equipment than mobile stations which is sensitive to radio power. These effects can be minimised e.g. with a correct choice of pulse ratio and ON periods.

As is obvious, embodiments of the invention are not limited to the embodiment described herein as an example, but they may vary in accordance with the scope of the appended claims.

Claims

1. Method of measuring traffic densities in a cellular radio network comprising at least mobile stations (MS) and base transceiver stations (BTS) and wherein during connections between base transceiver stations and mobile stations the mobile stations perform measurements of the quality of the radio channel and report their measurement results to the base transceiver station, characterized in that a test transmitter is located in the cell, such a test signal is sent from the test transmitter, which causes a change in the signal received from the cell's base transceiver station by mobile stations located within the service area of the test transmitter, and a mobile station is found to be within the service area of the test transmitter, if dependence on the test signal is detected in the measurement results reported by the mobile station.

2. Method as defined in claim 1, characterized in that the method determines the share of mobile stations located in the test transmitter's service area of all mobile stations served by the cell's base transceiver station.

3. Method as defined in claim 1, characterized in that the test signal is a periodic pulse signal.

4. Method as defined in claim 1, characterized in that the frequency band of the test signal covers one traffic frequency.

5. Method as defined in claim 1, characterized in that the frequency band of the test signal covers all traffic frequencies used in the area.

6. Method as defined in claim 1, characterized in that the analysis of dependence between quality measurements of mobile stations and the test signal is done through software.

7. Method as defined in claim 1, characterized in that the analysis of dependence between quality measurements of mobile stations and the test signal is done with a separate analyser and its control computer.

8. Method of defining a place for a new base transceiver station in a cellular radio network comprising at least mobile stations (MS) and base transceiver stations (BTS) and wherein during connections between base transceiver stations and mobile stations the mobile stations perform measurements of the radio channel quality and report their measurement results to the base transceiver station, characterized in that a test transmitter is located in the cell, such a test signal is sent by the test transmitter which causes a change in the signal received from the cell's base transceiver station by mobile stations located within the service area of the test transmitter, a mobile station is found to be within the service area of the test transmitter, if dependence on the test signal is found in the measurement results reported by the mobile station, and the place of the test transmitter is defined as the place for locating the new base transceiver station, if it is found that a sufficient number of the mobile stations served by the base transceiver station is located within the service area of the test transmitter.